public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed
* unknown type name 'TItype'
@ 2011-08-18 11:27 Maciej Bliziński
  2011-08-18 13:39 ` Jonathan Wakely
  0 siblings, 1 reply; 6+ messages in thread
From: Maciej Bliziński @ 2011-08-18 11:27 UTC (permalink / raw)
  To: gcc-help

Hello,

I'm continuing my work on the Solaris port of GCC-4.6.1.  Here's the
current status:

Solaris 9 sparc: builds fine
Solaris 9 i386: builds fine
Solaris 10 sparc: don't know yet
Solaris 10 i386: build fails

The sparc version is now building fine, but there is an issue when
building on Solaris 10 on the Intel architecture:

gmake[7]: Entering directory
`/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/objdir/i386-pc-solaris2.10/amd64/libgcc'
# If this is the top-level multilib, build all the other
# multilibs.
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/objdir/./gcc/xgcc
-B/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/objdir/./gcc/
-B/opt/csw/gcc4/i386-pc-solaris2.10/bin/
-B/opt/csw/gcc4/i386-pc-solaris2.10/lib/ -isystem
/opt/csw/gcc4/i386-pc-solaris2.10/include -isystem
/opt/csw/gcc4/i386-pc-solaris2.10/sys-include    -O2 -m32 -march=i386
-m64 -O2  -O2 -m32 -march=i386    -DIN_GCC   -W -Wall -Wwrite-strings
-Wcast-qual -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fPIC -g
-DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED
-fno-stack-protector   -I. -I. -I../../.././gcc
-I/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc
-I/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/.
-I/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc
-I/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../include
 -DHAVE_CC_TLS  -o fixtfti.o -MT fixtfti.o -MD -MP -MF fixtfti.dep
-fexceptions -c
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc/config/soft-fp/fixtfti.c
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc/config/soft-fp/fixtfti.c:34:1:
error: unknown type name 'TItype'
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc/config/soft-fp/fixtfti.c:34:8:
warning: no previous prototype for '__fixtfti' [-Wmissing-prototypes]
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc/config/soft-fp/fixtfti.c:
In function '__fixtfti':
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/libgcc/../gcc/config/soft-fp/fixtfti.c:38:3:
error: unknown type name 'UTItype'
(...)
gmake[7]: *** [fixtfti.o] Error 1
gmake[7]: Leaving directory
`/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/objdir/i386-pc-solaris2.10/amd64/libgcc'

Is this a known issue?

If anyone wanted to look more, the complete objdir archive is available here:
http://buildfarm.opencsw.org/~maciej/gcc-solaris-10-i386.tar.gz

Maciej

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: unknown type name 'TItype'
  2011-08-18 11:27 unknown type name 'TItype' Maciej Bliziński
@ 2011-08-18 13:39 ` Jonathan Wakely
  2011-08-18 14:15   ` Maciej Bliziński
  0 siblings, 1 reply; 6+ messages in thread
From: Jonathan Wakely @ 2011-08-18 13:39 UTC (permalink / raw)
  To: Maciej Bliziński; +Cc: gcc-help

2011/8/18 Maciej Bliziński :
>
> Is this a known issue?

It would help if you say what configure command you're using, I don't
want to download your archive to look.

i386-solaris is heavily tested and works for everyone else, so you
must be doing something unusual.

There are regular results posted to
http://gcc.gnu.org/ml/gcc-testresults/2011-08/ for both x86 and sparc
on solaris 2.9, 2.10 and 2.11

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: unknown type name 'TItype'
  2011-08-18 13:39 ` Jonathan Wakely
@ 2011-08-18 14:15   ` Maciej Bliziński
  2011-08-18 14:41     ` Jonathan Wakely
  0 siblings, 1 reply; 6+ messages in thread
From: Maciej Bliziński @ 2011-08-18 14:15 UTC (permalink / raw)
  To: Jonathan Wakely; +Cc: gcc-help

2011/8/18 Jonathan Wakely <jwakely.gcc@gmail.com>:
> 2011/8/18 Maciej Bliziński :
>>
>> Is this a known issue?
>
> It would help if you say what configure command you're using, I don't
> want to download your archive to look.
>
> i386-solaris is heavily tested and works for everyone else, so you
> must be doing something unusual.
>
> There are regular results posted to
> http://gcc.gnu.org/ml/gcc-testresults/2011-08/ for both x86 and sparc
> on solaris 2.9, 2.10 and 2.11

Interesting, is there an automated system that takes care of running
these tests and posting them to the mailing list?  Perhaps OpenCSW
could contribute.

Here is the configuration invocation that I used:

[===== NOW BUILDING: gcc-4.6.1 MODULATION isa-i386: ISA=i386 =====]
        [extract-modulated] complete for gcc.
        [patch-modulated] complete for gcc.
 ==> Running configure in work/solaris10-i386/build-isa-i386/gcc-4.6.1
cd work/solaris10-i386/build-isa-i386/gcc-4.6.1 && mkdir -p objdir &&
cd objdir && prefix="/opt/csw/gcc4" exec_prefix="/opt/csw/gcc4"
bindir="/opt/csw/gcc4/bin" sbindir="/opt/csw/gcc4/sbin"
libexecdir="/opt/csw/gcc4/libexec" datadir="/opt/csw/gcc4/share"
sysconfdir="/etc/opt/csw/gcc4" sharedstatedir="/opt/csw/gcc4/share"
localstatedir="/var/opt/csw/gcc4" libdir="/opt/csw/lib"
infodir="/opt/csw/gcc4/share/info"
lispdir="/opt/csw/gcc4/share/emacs/site-lisp"
includedir="/opt/csw/gcc4/include" mandir="/opt/csw/gcc4/share/man"
docdir="/opt/csw/gcc4/share/doc" sourcedir="/opt/csw/src"
CPPFLAGS="-I/opt/csw/gcc4/include" CFLAGS="-mtune=i686 -O2 -pipe -m32
-march=i386" CXXFLAGS="-mtune=i686 -O2 -pipe -m32 -march=i386"
LDFLAGS="-L/opt/csw/gcc4/lib/. -m32 -march=i386 -L/opt/csw/lib"
FFLAGS="-mtune=i686 -O2 -pipe -m32 -march=i386" FCFLAGS="-mtune=i686
-O2 -pipe -m32 -march=i386" ASFLAGS="" OPTFLAGS="-mtune=i686 -O2 -pipe
-m32 -march=i386" CC="/opt/csw/gcc4/bin/gcc"
CXX="/opt/csw/gcc4/bin/g++" CC_HOME="/opt/csw/gcc4" CC_VERSION="gcc
version 4.3.3 (GCC) " CXX_VERSION="gcc version 4.3.3 (GCC) "
GARCH="i386" GAROSREL="5.10" GARPACKAGE="bootstrap-4.6"
LD_OPTIONS="-R/opt/csw/gcc4/lib -R/opt/csw/lib/\$ISALIST
-R/opt/csw/lib" CFLAGS_FOR_TARGET="-O2 -m32 -march=i386   "
CXXFLAGS_FOR_TARGET="-O2 -m32 -march=i386   "
PKG_CONFIG_PATH="/opt/csw/lib/pkgconfig"
DESTDIR="/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/install-isa-i386"
/home/maciej/src/opencsw/pkg/gcc4/branches/bootstrap-4.6/work/solaris10-i386/build-isa-i386/gcc-4.6.1/configure
--with-pkgversion=OpenCSW --prefix=/opt/csw/gcc4
--exec_prefix=/opt/csw/gcc4 --bindir=/opt/csw/gcc4/bin
--sbindir=/opt/csw/gcc4/sbin --libexecdir=/opt/csw/gcc4/libexec
--datadir=/opt/csw/gcc4/share --sysconfdir=/etc/opt/csw/gcc4
--sharedstatedir=/opt/csw/gcc4/share --localstatedir=/var/opt/csw/gcc4
--libdir=/opt/csw/lib --infodir=/opt/csw/gcc4/share/info
--includedir=/opt/csw/gcc4/include --mandir=/opt/csw/gcc4/share/man
--exec-prefix=/opt/csw/gcc4 --with-gnu-as --with-as=/opt/csw/bin/gas
--without-gnu-ld --with-ld=/usr/ccs/bin/ld --enable-nls
--with-included-gettext --with-libiconv-prefix=/opt/csw --with-x
--with-mpfr=/opt/csw --with-gmp=/opt/csw --enable-java-awt=xlib
--enable-libssp --enable-objc-gc --enable-threads=posix
--enable-stage1-languages=c --enable-languages=c,c++,fortran,java,objc
--enable-bootstrap --with-system-zlib --with-cpu-32=i386
--with-cpu-64=x86_64

My working hypothesis is that the --with-cpu-64=x86_64 setting
confused the ifdef blocks.  I'm running a build with just
--with-cpu-32=i386, I'll see how it goes.  Do you notice anything
unusual in the above invocation?

Maciej

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: unknown type name 'TItype'
  2011-08-18 14:15   ` Maciej Bliziński
@ 2011-08-18 14:41     ` Jonathan Wakely
  2011-08-21  8:31       ` Maciej Bliziński
  0 siblings, 1 reply; 6+ messages in thread
From: Jonathan Wakely @ 2011-08-18 14:41 UTC (permalink / raw)
  To: Maciej Bliziński; +Cc: gcc-help

2011/8/18 Maciej Bliziński:
>
> Here is the configuration invocation that I used:
>
> [===== NOW BUILDING: gcc-4.6.1 MODULATION isa-i386: ISA=i386 =====]
>        [extract-modulated] complete for gcc.
>        [patch-modulated] complete for gcc.
>  ==> Running configure in work/solaris10-i386/build-isa-i386/gcc-4.6.1
> cd work/solaris10-i386/build-isa-i386/gcc-4.6.1 && mkdir -p objdir &&
> cd objdir &&

Building in a sub-directory of the sources is NOT supported.

http://gcc.gnu.org/install/configure.html

Also, I fell asleep halfway through reading all those variables
assignments, why don't you try keeping it simple just to see if you
can build a basic configuration, then start adding all those tweaks if
you really think you need them

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: unknown type name 'TItype'
  2011-08-18 14:41     ` Jonathan Wakely
@ 2011-08-21  8:31       ` Maciej Bliziński
  0 siblings, 0 replies; 6+ messages in thread
From: Maciej Bliziński @ 2011-08-21  8:31 UTC (permalink / raw)
  To: Jonathan Wakely; +Cc: gcc-help

2011/8/18 Jonathan Wakely <jwakely.gcc@gmail.com>:
> 2011/8/18 Maciej Bliziński:
>>
>> Here is the configuration invocation that I used:
>>
>> [===== NOW BUILDING: gcc-4.6.1 MODULATION isa-i386: ISA=i386 =====]
>>        [extract-modulated] complete for gcc.
>>        [patch-modulated] complete for gcc.
>>  ==> Running configure in work/solaris10-i386/build-isa-i386/gcc-4.6.1
>> cd work/solaris10-i386/build-isa-i386/gcc-4.6.1 && mkdir -p objdir &&
>> cd objdir &&
>
> Building in a sub-directory of the sources is NOT supported.
>
> http://gcc.gnu.org/install/configure.html

Thanks for spotting this.  I adjusted the build description to use a
directory beside the sources.

> Also, I fell asleep halfway through reading all those variables
> assignments, why don't you try keeping it simple just to see if you
> can build a basic configuration, then start adding all those tweaks if
> you really think you need them

I went this general path.  After some experimentation, I found out
that CFLAGS and CFLAGS_FOR_TARGET are only necessary for the sparc
architecture.  When building for i386, it's best to unset CFLAGS (and
friends, including FFLAGS and FCFLAGS).

Usually, I try to keep package build descriptions for sparc and i386
as similar as possible.  In this case, I had to have them diverge a
bit.  Thanks again for the help, I got GCC to build on the Intel
architecture and will proceed with testing.

Maciej

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: unknown type name 'TItype'
@ 2011-08-18 14:18 Dennis Clarke
  0 siblings, 0 replies; 6+ messages in thread
From: Dennis Clarke @ 2011-08-18 14:18 UTC (permalink / raw)
  To: Maciej Bliziński; +Cc: Jonathan Wakely, gcc-help


> 2011/8/18 Jonathan Wakely <jwakely.gcc@gmail.com>:
>> 2011/8/18 Maciej Bliziński :
>>>
>>> Is this a known issue?
>>
>> It would help if you say what configure command you're using, I don't
>> want to download your archive to look.
>>
>> i386-solaris is heavily tested and works for everyone else, so you
>> must be doing something unusual.
>>
>> There are regular results posted to
>> http://gcc.gnu.org/ml/gcc-testresults/2011-08/ for both x86 and sparc
>> on solaris 2.9, 2.10 and 2.11

And Solaris 8 ( sparc and i386 ) which is on contract support until spring
next year.


-- 
--
http://pgp.mit.edu:11371/pks/lookup?op=vindex&search=0x1D936C72FA35B44B
+-------------------------+-----------------------------------+
| Dennis Clarke           | Solaris and Linux and Open Source |
| dclarke@blastwave.org   | Respect for open standards.       |
+-------------------------+-----------------------------------+

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2011-08-21  8:31 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-08-18 11:27 unknown type name 'TItype' Maciej Bliziński
2011-08-18 13:39 ` Jonathan Wakely
2011-08-18 14:15   ` Maciej Bliziński
2011-08-18 14:41     ` Jonathan Wakely
2011-08-21  8:31       ` Maciej Bliziński
2011-08-18 14:18 Dennis Clarke

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).